1

HYDROCRACKING OF LONG CHAIN LINEAR PARAFFINS

Ilenia Rossettia*, Chiara Gambarob, Vincenzo Calemmab

a Dip. Chimica Fisica ed Elettrochimica, Università degli Studi di Milano,

via C. Golgi 19, I-20133 Milano, Italy

b ENI S.p.A. R&M Division, via F. Maritano 26, I-20097 S. Donato Milanese, Italy

ABSTRACT

The hydrocracking reactivity of two model compounds, namely n-C16H34 (n-C16) and n-

C28H58 (n-C28), was investigated on a Pt/SiO2-Al2O3 catalyst. Conversion and products

distribution have been determined under a wide range of operating conditions (i.e. pressure:

20-80 bar; temperature: 270-330°C; weight hourly space velocity: 0.33-1.0 h-1; H2/n-paraffin

feeding ratio (0.05-0.15 wt/wt). The latter were changed according to a central composite

design. The present paper summarises the results obtained on both the model paraffins,

depending on the reaction conditions. A first, simple kinetic elaboration is also presented,

based on an ideal PFR model and a first order kinetics. The reaction confirmed to be first

order with respect to the n-paraffin. Experimental data showed that for both n-C16 and n-C28

conversion was affected by H2/n-paraffin ratio. The change of conversion was explained in

terms of vapour liquid equilibrium (VLE), which in turn is affected by the H2/n-paraffin ratio, so

leading to a different vaporisation degree of reactant. In agreement with the VLE data, the

effect of H2/n-paraffin on conversion was lower for n-C28. VLE calculations have been carried

out to estimate the H2 partial pressure and degree of vaporisation of the normal paraffin. The

reaction order for hydrogen was -1 and -0.5 for n-C16 an n-C28, respectively. However, in the

case of n-C16 the data obtained at the lower bound of the pressure range examined

displayed an increase of the reaction order. The apparent activation energy was calculated

after correction of the contact time taking into account the liquid-vapour equilibrium: similar

values have been estimated for n-C16 and n-C28, ca. 32 and 31 kcal/mol, respectively.

Keywords: Hydrocracking; hydroisomerisation; upgrading of Fischer-Tropsch waxes.

* Corresponding author: fax +39-02-50314300, e-mail ilenia.rossetti@unimi.it

2

1 - INTRODUCTION

Hydroisomerisation and hydrocracking of alkanes are important refinery processes to

obtain high-octane-number reformulated gasoline. Furthermore, waxes, inducing poor cold-

flow properties, have to be eliminated through dewaxing and isomerisation of the long-chain

n-alkanes to moderately branched alkanes, the latter being excellent components for lube

oils. The hydroconversion of the synthetic wax produced by the Fischer-Tropsch (FT) process

represents a different route for the production of high quality fuels [1]. The FT products are

essentially a mixture of linear paraffins spread over a wide range of molecular weight (C1-

C70), due to the polymerisation nature of the reaction. Cobalt catalyzed low temperature FT

process is characterised by a products distribution strongly shifted towards heavy

components [2,3]. An efficient way to maximize the yields in the middle distillate cut is to

subject the FT products to an hydrocracking stage [4,5].

The reaction is efficiently carried out over bi-functional catalysts, consisting of a metal,

responsible for hydro-dehydrogenation reactions, dispersed on an oxide support, whose acid

sites are responsible of C-C and C-H bonds activation via a carbocationic mechanism.

Various metals (e.g. Pt, Ni, Mo) and different acidic supports, such as zeolites, SAPO,

sulphated zirconia and amorphous silica-alumina, have been employed [1]. The degree of

isomerisation vs. cracking can be modulated by proper selection of the acid strength or by

taking advantage of the shape selectivity of some zeolite-structured materials. For example,

strong acidity (zeolites) favours cracking, while moderate acidity (SAPO) shows higher

selectivity for isomerisation [6]. However, catalyst coking and lower selectivity may result from

an improper balance between acid sites strength and hydrogenation activity [7]. Some

mesoporous materials have been investigated as well, though they seemed suitable for

hydroisomerisation more than for cracking, due to milder acidity and lower thermal resistance.

With Pt loaded on sulphated zirconia, the cracking/isomerisation ratio rapidly increases with

the carbon chain length of the reactant molecule [8].

As for the metallic function, the easier the hydrogenation step, the lower is the

carbocation lifetime, with subsequent lower probability of secondary cracking events [1]: for

the hydrocracking of “clean” feedstocks (free of sulphur, metals and oxygenated compounds)

noble metals are suitable hydrogenating active sites, being very active; otherwise, also non-

noble transition metals, such as Ni, Co, Mo and W, are used [6].

The hydrocracking mechanism can be schematically represented as a series of

isomerisation and cracking steps: the n-paraffin in the feed is isomerised to the corresponding

3

mono-branched paraffin, which is in turn converted to the di-branched paraffin and so on; the

multi-branched paraffins are then cracked to lighter products [9].

The kinetics of the reaction has been modelled following different approaches. A

fundamental one is based on the Langmuir-Hinshelwood-Hougen-Watson approach (LHHW),

which decomposes the reaction in a series of elementary steps, to be modelled with proper

rate expressions and subsequently lumped to describe the products distribution [10,11]. A

different approach is represented by the “single event kinetics” (SE): a network of elementary

reactions is designed taking into account the formation of each single component of the

product mixture [12-14]. The SE kinetics is based on the “chemical knowledge” of each

elementary step involved in the reaction; thus rate parameters have a general meaning and

can be applied to different feeds [15]. On the other hand, due to the complicated network of

rate expressions, the SE approach is best applied to model compounds rather than to

complex mixtures, such as FT waxes, due to extremely increasing calculation time, although

efforts have been made to simplify some points [14] and to extrapolate the results to real

mixtures. The kinetic description of the hydrocracking of complex mixtures has been generally

achieved, at least up to now, by means of lumped models [16-20].

According to our knowledge, in literature there are very few kinetic studies carried out on

single components of relatively high molecular weight [9,21]. Such data are indeed needed to

better support the interpretation of the behaviour of complex mixtures. The aim of the present

work was then to collect kinetic data for the hydrocracking of two paraffins of relatively high

molecular weight, namely n-C16H34 (n-C16) and n-C28H58 (n-C28). The effect of four different

parameters, i.e. temperature, pressure, H2/n-paraffin ratio and contact time, on conversion

and selectivity has been checked following a factorial experimental plan. A first simple kinetic

analysis is also presented.

2 - EXPERIMENTAL

2.1 – Set up of the apparatus

The experimental apparatus was made up of a downflow continuous Incoloy 800 tubular

reactor (400 mm in length, 10.3 mm internal diameter), surrounded by two heavy AISI 316

hemi-cylindrical blocks and heated by an electric furnace, controlled by an Eurotherm (mod.

822, coupled to a mod. 425S power unit) TIC. The paraffins, stored in a heated tank, so to

keep them in liquid form, were fed to the reactor by means of a heated syringe pump (Isco,

4

mod. 314). H2 flow rate was regulated by means of a Brooks (mod. 5850S) mass flow meter,

governed by a control unit (Brooks, mod.154). A pressurised storage tank at the exit of the

reactor was employed for the collection of liquid products, whereas the flow rate of the

gaseous fraction was measured by a volumetric wet-test meter (Elster Handel, ser.

63004054). The reaction pressure was regulated by means of a home-made PTFE-

membrane relief valve, pressurised with nitrogen at the desired value. A high-precision

manometer, a bursting-disc safety device and heating devices on the whole feeding and

reaction lines completed the apparatus.

Kinetic data have been collected on ca. 5 g of a Pt/amorphous SiO2-Al2O3 catalyst in 20-

45 mesh particle size (bed volume ca. 10 cm3 and bed length ca. 12 cm, fitting the isothermal

central part of the furnace). Catalyst wetting has been checked for this particle size and under

the selected operating conditions according to the Gierman and Harmsen criteria as suggested

in [22]. The void part of the reactor, above and below the catalyst bed, was filled with quartz

beads (10-20 mesh), previously washed with diluted HCl and calcined at 450°C in air.

Catalyst activation was carried out by heating up to 400°C in flowing H2.

Catalytic activity has been monitored for 5 days until the attainment of steady-state

conditions for every new feed. After completion of the kinetic experimentation on n-C16 the

whole apparatus was carefully washed at first with n-C7, then with n-C28, and again the

activity was monitored until the attainment of steady state conversion before starting the

collection of the data on n-C28.

2.2 - Planning of experimental tests

Kinetic data have been collected using a central composite design approach [23], based

on four variables: temperature (T), pressure (P), weight hourly space velocity (WHSV), H2/n-

paraffin (wt/wt) and for every set of four parameters conversion of the reactant (n-C16 or n-

C28) and selectivity to the various products or lumps have been determined.

5

The central run conditions (level [0 0 0 0]) have been set in order to attain ca. 50%

conversion of the n-paraffin. The experiments have been made on two levels [+1 -1], plus for

every factor four further levels were added [+2 +1 -1 -2] and central runs were periodically

and frequently repeated. The complete experimental plan for each feed included a first set of

25 runs, made in a random sequence, plus various additional tests to address specific topics

and for reproducibility checking for a total of ca. 50 runs. The values of the four factors

corresponding to each level are listed in Table 1. After the completion of kinetic data

measurements on n-C28, some tests have been added for n-C16, showing reproducibility

with previous runs and excluding any severe deactivation of the catalyst.

2.3 – Analysis of the reaction products

For every run a one hour-on-stream time was left to attain steady state conditions. This

time interval showed sufficient for this purpose during the preliminary tests carried out for

setting up the apparatus. Then, a 1 hour run started for each test, during which the liquid

products were collected in the pressurised tank and subsequently analysed, while effluent gas

analysis was carried out at least twice during each run. Aliquots of the collected liquid

products were then dissolved in CS2 and analysed.

The analysis of both liquid and gaseous products has been made off-line by means of a

properly calibrated Agilent 6890 gas chromatograph, equipped with flame ionisation (FID) and

hot-wire (HWD) detectors and with HP-5 and PoraPlot Q columns for liquid and gas analysis,

respectively, with H2 as carrier gas. Due to the relatively poor resolution for different isomers,

the results have been lumped as linear (PC), mono-methyl substituted (MC), di-methyl

substituted (DC), tri-methyl substituted (TC) and ethyl-substituted (EC) paraffins, as

exemplified in Fig.1.

3 - RESULTS AND DISCUSSION

3.1 – Effect of the main operating parameters on activity

6

A typical distribution of hydrocracking products as a function of chain length is presented

in Fig.1a, which reports the average results of the central runs carried out on n-C28. The

cracking products consist mainly of linear paraffins and mono-branched compounds, followed

in descending order by di-, tri-branched alkanes. Moreover, experimental results (not shown

for the sake of brevity) display that the molar ratio between iso- and n-paraffins increases with

conversion, at constant temperature, in a rather exponential way, due to consecutive

isomerisation reactions [24]. Monobranched isomers are essentially a mixture of monomethyl

(predominant) and ethyl-paraffins. Formation of methyl branching can be readily explained

according to the protonated cyclopropane mechanism (PCP) [25]. Based on the same

reaction mechanism, ethyl-substituted isomers are less favoured, due to the lower stability of

the cyclobutane intermediate [26].

A flat products distribution was observed at low conversion, which evolves to a bell-

shaped curve at high conversion, as reported in Fig. 1b for n-C16. A shift towards low-

molecular weight products was observed particularly at high conversion levels, where the

subsequent cracking of first formed products is quantitatively significant. As previously

reported in literature for lower molecular weight normal paraffins, regardless the operating

conditions [27,28] the isomer and cracking yields, reported in figures 2a and 2b, are a unique

function of the total conversion, thus indicating a similar activation energy of the

hydroisomerisation and hydrocracking reactions. Furthermore, the evolution of both systems

is in agreement with a reaction pathway where the conversion of n-alkane occurs through a

series of consecutive reactions: the n-alkane is first isomerised into mono-branched isomers,

which undergo subsequent isomerisation steps and cracking reactions.

The effect of the operating conditions, temperature, pressure, H2/n-paraffin ratio and

WHSV, on the conversion of the two model compounds (n-C16 and n-C28) is reported in

Fig.3 a, b, c and d. Fig.3a depicts the dependence of conversion on temperature for both

reactants. Conversion always increased with temperature, as expected, due to kinetic

7

reasons. The conversion of n-C28 was always higher than that of n-C16, this can be

accounted for by the fact that the reactivity is expected to be proportional to the number of

secondary carbon atoms per molecule or, as proposed by Sie, to Cn-6 and Cn-4 for

hydrocracking and hydroisomerisation respectively [29]. However, as reported in several

works [30,31], the higher reactivity of heavier n-paraffins can be also ascribed to their

stronger physisorption, which lead to a higher density on the catalyst surface and

consequently to higher reaction rates.

Fig.3b reports the effect of total pressure on conversion. For both feedstocks conversion

decreased with increasing pressure. The negative dependence of normal alkane conversion

on hydrogen can be explained in terms of a bifunctional mechanism where the first step is the

formation of olefins at the metal site, subsequent formation of secondary carbenium at the

acidic site and then the various steps leading to the final products [11]. In this scheme the rate

determining step is the rearrangement of the secondary carbenium while the other reactions

are considered to be in quasi equilibrium conditions, hence an increase of hydrogen pressure

will lead to lower steady state concentrations of n-alkenes and consequently of secondary

carbenium ions, which results in a decrement of the limiting step rate [32,33]. This conversion

vs. pressure trend also confirms that the catalyst is operating under “ideal” conditions, i.e. it is

possible to consider the hydro/dehydrogenation steps in quasi-equilibrium [34,35].

Feed conversion vs. H2/n-paraffin weight ratio is reported in Fig.3c. Higher H2/n-paraffin

inlet ratios led to an increase in the paraffin conversion. The effect is much more pronounced

in the case of n-C16 than for n-C28. Notably, the reaction mechanism foresees a decrease in

the reaction rate with the increase of the H2/n-paraffin ratio rather than an increase as

experimentally observed [12]:

2

2

/1

/

,

,

HnPPROTDH

PLps

nPPL

HnPPROTDH

totiso

PPKKKc

PK

PPKKckr

(1)

8

where kiso is the kinetic constant for the isomerisation reaction, KDH and KPROT are the hydro-

dehydrogenation and protonation equilibrium constants, KL,P is the Langmuir physisorption

equilibrium constant for the n-paraffin, ctot is the concentration of acid sites, cps is the

maximum concentration of paraffin in the catalyst pores, Pi is the partial pressure of the

reactant (H2 or paraffina). The equation refers to the case of low conversion, where the iso-

paraffin concentration is low. The promoting effect of the H2/wax ratio on conversion should

rather be explained considering the role of this parameter on vapour-liquid equilibrium (VLE)

[37]. Depending on the reaction conditions and paraffin used during the tests, a fraction of the

normal paraffin evaporates from the liquid to the vapour phase. The “degree of evaporation”

of the hydrocarbon changes with the H2/wax ratio, as demonstrated by calculations carried

out with a routine based on SRK equation of state, considered one of the most suitable for

this problem [38,39], whose results are reported in figure 3c. The same calculation showed

that only a small fraction of H2 (ca. 0.5 wt% with respect to the H2 fed, in first approximation

considered negligible) was dissolved in liquid phase in the case of n-C16 under the conditions

of the central run, taken as reference. We observed that in both cases the paraffin fraction in

the vapour phase increases along with the increase of the H2/n-paraffin ratio, but remarkable

quantitative differences exist as for the entity of the phenomenon. In the case of n-C16 the

fraction in vapour phase at the reactor inlet ranges between 75.1 and 93.1%, whereas in the

case of n-C28, its percentage in gas phase is much lower and displays a lower variation. In

these circumstances, the increase in conversion with the rise of H2/n-paraffin ratio can be

rather ascribed to the change in VLE, leading to higher fraction of reactant and products in the

gas phase. We can assume that the reaction occurs in an ideal trickle flow regime and that

the reactive part is in liquid phase, in contact with the catalyst, the gas phase simply being in

equilibrium with the liquid one. On this basis, the conversion would in first approximation be a

a In this specific case, due to their high molecular weight, the paraffins are not in gas phase, so the notation Pn-P

is not completely correct: a more reliable way to write this rate expression for heavy hydrocarbons is to use

fugacity [36]

9

function of the space velocity of the liquid phase, rather than of the overall space velocity at

the rector inlet. The equation defining the relationship between conversion and residence time

for a first order reaction: yf = 1 - e-k/WHSV.The plot of ln(1-yf) vs. 1/WHSV (or ) should be a

straight line, whose slope is the kinetic constant k. The latter is only affected by temperature

and should remain constant when other operating conditions are changed. In the case of n-

C16, different values of k can be estimated changing the H2/n-paraffin ratio when the inlet

WHSV=2 h-1 is considered (see figure 4a). This happens because a significant amount of the

reactant evaporates: if the real space velocity of the liquid phase, WHSVL=(1-f).WHSV (where

f is the degree of evaporation) is considered, then the points at different H2/n-paraffin ratios

are aligned on the same trace, as shown in figure 4b. The minimum x-value that the plot can

reach is L = 0.5 h, that is the case in which the paraffin remains all in liquid phase and the

space velocity of the liquid phase coincides with the inlet WHSV. The point where the plot

crosses the line L = 0.5 h gives the conversion when no evaporation occurs, and is related to

the intrinsic reactivity of the liquid phase.

Finally, the effect of WHSV on conversion is shown in Fig.3d. As expected, conversion

decreased with increasing space velocity, i.e. with decreasing contact time. The cracking

products concentration increased rather monotonously with conversion, as the multi-branched

C16 or C28 isomers. By contrast, the mono-substituted isomers concentration increased with

conversion and contact time, decreasing at the lowest value of WHSV, predominantly due to

transformation into cracking products. It is worth noticing that the products distribution shifted

towards lighter products with increasing contact time, due to secondary cracking for both

feedstocks.

Summarising, the dependence of conversion on the above reported parameters can be

sketched as follows: i) the higher the conversion, the lower is the amount of intermediate

molecular weight products, indicating the presence of consecutive cracking reactions; ii) a

temperature increase brings about an increase of conversion, due to kinetic reasons; iii)

10

conversion decreases with increasing pressure, due to the increase of hydrogen partial

pressure; iv) an increase of WHSV means a decrease of contact time and has a negative

effect on conversion; v) the H2/n-paraffin ratio affects the liquid-vapour equilibrium of the

reacting mixture. A high H2/n-paraffin ratio (i.e. higher hydrogen concentration) increases the

degree of vaporisation of light species and causes a higher concentration of the heavier

species in the liquid phase. The latter being more reactive, the conversion increases. This

effect was much more evident for n-C16 than for n-C28.

3.2 – Simplified kinetic analysis

A simplified approach was followed to estimate the reaction order with respect to the

paraffin (n-C16 or n-C28) and H2 and to calculate the apparent activation energy. For this

purpose, the data were analyzed considering an ideal plug-flow reactor and a first order rate

expression with respect to the paraffin. In order to check the reliability of the latter hypothesis,

at first the hydrogen partial pressure was assumed constant, H2 conversion being usually

lower than 10%. With such constraints, the following relationship holds between conversion

and contact time:

- ln (1-yf) = k * (2)

yf being conversion, k the kinetic constant and the contact time. If the hypothesis of first

order kinetics for the paraffin is correct, by plotting Y = - ln (1-yf) vs. , a straight line, with zero

intercept and slope k should be obtained.

Under reaction conditions the feed is subjected to vapour-liquid equilibrium, the contact

time in the simplified model should actually be that of the liquid phase covering the catalyst

particle and considered as the reactive portion of the paraffin, the contribution to reactivity of

the gas phase being less important (vide supra). The tests performed at different space

velocities were used for this purpose and the results are plotted in Fig.5: the simplified model

seems to work well, since good correlations have been obtained for both the paraffins.

11

This approach was applied to the data collected at different temperatures: the estimated

kinetic constants were used to calculate the apparent activation energies for the two reactants

by means of the Arrhenius equation: regression of the results gave a value of 32.1 kcal/mol

for n-C16 and 30.8 kcal/mol for n-C28. These are roughly in line with literature data based on

lumped kinetic models, with a slight overestimation in some cases [9,40,41] and support the

observed higher reactivity of n-C28 with respect to n-C16.

3.3 - Reaction order with respect to H2

The reaction order with respect to H2 was evaluated introducing a power law rate

expression and, as a first approximation, neglecting the H2 amount dissolved in liquid phase.

Furthermore, pH2 has been calculated by subtraction of the paraffin partial pressure

(calculated on the basis of VLE) from the total pressure. The different composition along the

reactor has not been taken into account. Hence, the following rate equation was used,

considering the partial pressure of the n-paraffin as a constant:

Reaction rate (mol/h) = k’ * pH2 (3)

Linearization has been achieved by a bi-logarithmic plot of the reaction rate against

hydrogen pressure: the slope of the regression line corresponds to the H2 reaction order, .

Results are reported in Fig.6: a linear plot was obtained for the case of n-C28, with a

slope of -0.55. Notably, H2 partial pressure is close to the total pressure, in line with a low

volatility of the paraffin under reaction conditions. By contrast, n-C16 partial pressure was not

negligible, in particular at the lowest pressures. Even if the correction due to VLE was applied,

a non linear pattern has been obtained with n-C16, as depicted in figure 6. Then only the

three points at higher pressure have been taken into account for regression and a reaction

order for H2 of -1.0 was estimated for n-C16. A negative reaction order was indeed expected,

due to the first dehydrogenation step of the reaction [10,42]. However, so different values for

the two paraffins seem surprising, as well as the singular pattern obtained with n-C16.

12

Ribeiro et al. [41] proposed an explanation for the fractional H2 reaction order, in light of

the bifunctional mechanism usually assumed. Removing the adsorption-desorption

equilibrium expression from equation (1), it can be rearranged as follows:

*

2

*

21

21

nPH

nPtotiso

cKKP

cKKckr

(4)

where cnP* is the concentration of the adsorbed n-paraffin, kiso the kinetic constant of

isomerisation, K1 the equilibrium constant of dehydrogenation, K2 that of carbenium ion

formation, while ctot is given by the product of K1,K2 and the concentration of acid sites.

Looking at eq (4), if the product *21 nPcKK is negligible with respect to the H2 partial

pressure, the reaction order with respect to hydrogen is actually -1, as in the case of light

paraffins (C6 [41], C10 [10] and present data on n-C16), otherwise a fractional reaction order

is found, as in our case with n-C28. In fact, K1 and K2 increase with the chain length for

entropic reasons, as proposed by de Gauw and co-workers [43]. In addition adsorption

phenomena increase with chain length. As a consequence, the product *21 nPcKK is never

negligible with respect to PH2 in the case of n-C28. Otherwise, in the case of n-C16 the trend

observed witnesses the transition between two regimes: from high hydrogen pressure values,

where the term *21 nPcKK is negligible with respect to the H2 partial pressure, to low values

of the latter where the term *21 nPcKK is no more negligible, thus leading to a fractional

order. A factor which may contribute in this direction is the higher degree of evaporation at

low pressure, as shown in Figure 6. Higher evaporation degrees correspond to a higher

paraffin partial pressures and therefore to lower H2 partial pressures: under these conditions,

the product *21 nPcKK is not negligible and a fractional reaction order towards H2 is obtained.

4 - CONCLUSIONS

13

A set of kinetic data for the hydrocracking reaction has been collected on n-C16H34 and

n-C28H58. The effect of the main reaction parameters (temperature, pressure, H2/wax ratio and

space velocity) on conversion and products distribution has been investigated. Conversion of

both normal paraffins is affected by the H2/n-paraffin ratio indicating that a significant role is

played by the VLE. In agreement with the VLE data, the effect of H2/n-paraffin ratio on

conversion is higher in the case of n-C16, which displays a wider change of vaporization

degree. Both n-C16 and n-C28 showed a first reaction order with respect to the n-paraffin.

Reaction order with respect to hydrogen resulted to be ca. -1 for n-C16 and ca. - 0.5 for n-

C28. However, different regimes can be envisaged in the case of n-C16, depending on

hydrogen partial pressure. Tentatively, we suggest that the fractional order for hydrogen

observed for n-C28 is due to a term in the rate equation containing the constant of protonation

and dehydrogenation, which is not negligible. Within the same line of reasoning one can

explain the change of reaction order with respect to hydrogen displayed by the n-C16. In

agreement with the values generally reported in literature [9,10] the apparent activation

energies were ca. 32.1 and 30.8 kcal/mol, for n-C16 and n-C28 respectively.

ACKNOWLEDGMENTS

The valuable help of A. Landoni, M. Bos and B. Cortese is gratefully acknowledged.

REFERENCES

1. V.M. Akhmedov, S.H. Al-Khowaiter, Catal. Rev., 49 (2007) 33

2. M.E. Dry, Catal. Today, 71 (2002) 227

3. R. L. Espinoza, A. P. Steynberg, B. Jager, A. C. Vosloo, Appl. Catal. A: General, 186

(1999) 13

4. M.J. Murphy, J.D. Taylor, R.I. Mc Cormic, Compendium of Experimental Cetane

Number Data, NREL/SR-540-36805 (2004)

5. Handbook of Chemistry and Physics, D.R. Lide, Ed., 85th Edition, 2004-2005, CRC

Press

14

6. H. Deldari, Appl. Catal. A:Gen 293 (2005) 1

7. J.W. Thybaut, C.S. Laxmi Narasimhan, J.F. Denayer, G.V. Baron, P.A. Jacobs, J.A.

Martens, G.B. Marin, Ind. Eng. Chem. Res. 44 (2005) 5159

8. E. Iglesia, S.L. Soled, G.M. Kramer, J. Catal., 144 (1993) 238

9. V. Calemma, S. Peratello, C. Perego, Appl. Catal. A 190 (2000) 207

10. M. Steijns, G. F. Froment, Ind. Eng. Chem. Prod. Res. Dev., 20 (1981) 660

11. M.A. Baltanas, H. Vansina, G. F. Froment, Ind. Eng. Chem. Prod. Res. Dev., 22 (1983)

531

12. G.F. Froment, Catal Today, 1 (1987) 455

13. G.F. Froment, Catal. Rev. Sci. Eng., 47 (2005) 83

14. E. Valéry, D. Guillaume, K. Surla, P. Galtier, J. Verstraete, D. Schweich, Ind. Eng.

Chem. Res., 46 (2007) 4755

15. G.D. Svoboda, E. Vynckier, B. Debrabandere, G.F. Froment, Ind. Eng. Chem. Res., 34

(1995) 3793

16. N. Bhutani, A.K. Ray, G. P. Rangaiah, Ind. Eng. Chem. Res., 45 ( 2006) 1354

17. S. Sànchez, J. Ancheyta, Energy & Fuels, 21 (2007) 653

18. L. Pellegrini, S. Bonomi, S. Gamba, V. Calemma, D. Molinari, Chem. Eng. Sci., 62

(2007) 5013

19. P. Castaño, J.M. Arandes, B. Pawelec, M. Olazar, J. Bilbao, Ind. Eng. Chem. Res., 47

(2008) 1043

20. P. Balasubramanian, S. Pushpavanam, Fuel, 87 (2008) 1660

21. V. Calemma, S. Peratello, F. Stroppa, R. Giardino, C. Perego, Ind. Eng. Chem. Res.,

43 (2004) 934

22. L.F. Ramìrez, J. Escobar, E. Galvàn, H. Vaca, F.R. Murrieta, M.R.S. Luna, Petroleum

Science and Technology, 22 (2004) 157

23. K. Hinkelmann, O. Kempthorne, Design and analysis of experiments, Vol. 2: Advanced

experimental design, Wiley, 2005.

24. T.D. Pope, J.F. Kriz, M. Stanciulescu, J. Monnier, Appl. Catal. A, 233 (2002) 45

25. S.T. Sie, Ind Eng. Chem. Res., 31 (1992) 1881

26. C. Marcilly, Catalyse acido-basique. Application au raffinage et à la pétrochimie,

Editions Technip 2003, vol.1, ch.3, p. 168

27. M. Steijns, G. F. Froment, P. Jacobs, J. Uytterhoeven, J. Weitkamp, Ind. Eng. Chem.

Prod. Res. Dev., 20 (1981) 654

28. J. Weitkamp, A. Jacobs, J.A. Martens, Appl. Catal. 8 (1983) 123

29. S.T. Sie, Ind Eng. Chem. Res., 32 (1993) 403

15

30. J.F.M. Denayer, B. De Jonckheere, M. Hioch, G.B. Marin, G. Vanbutsele, J.A. Martens,

G.V. Baron, J. Catal., 210 (2002) 445

31. J.F.M. Denayer, R.A. Ocakoglu, W.Huybrechts, B. De Jonckheere, P. Jacobs, S. Calero,

R. Krishna, B. Smit, G.V. Baron, J.A. Martens, J. Catal., 220 (2003) 66

32. M.A. Baltanas, K.K. van Raemdonck, G.F. Froment, S.R. Mohedas, Ind. Eng. Chem.

Res. 28 (1989) 899

33. H. Kumar, G.F. Froment, Ind. Eng. Chem. Res., 46 (2007) 4075

34. J.W. Thybaut, C.S. Laxmi Narasimhan, J.F. Denayer, G.V. Baron, P.A. Jacobs, J.A.

Martens, G.B. Marin, Ind. Eng. Chem. Res., 44 (2005) 5159

35. J.W. Thybaut, C.S. Laxmi Narasimhan, G.B. Marin, Catal. Today, 111 (2006) 94

36. L.A. Pellegrini, S. Gamba, V. Calemma, S. Bonomi, Chem. Eng. Sci. 63 (2008) 4285

37. V. Calemma, S. Correra, C. Perego, P. Pollesel, L. Pellegrini, Catal. Today 106 (2005)

282

38. "Equations of State for Fluids and Fluid Mixtures", J.V. Sengers, R.F. Kayser, C.J.

Peters, H.J. White, Jr., eds., Elsevier, Amsterdam, 2000.

39. J.O. Valderrama, Ind. Eng. Chem. Res., 42 (2003) 1603.

40. B. Debrabandere, G.F. Froment Stud.Surf.Sci.Catal. 106 (1997) 379

41. F. Ribeiro, C. Marcilly, M. Guisnet, J.Catal. 78 (1982) 267

42. V.M. Akhmedov, K.J. Klabunde, J. Mol. Catal., 45 (1988) 193

43. F.J.M.M. de Gauw, J. van Grondelle, R.A. van Santen, J. Catal. 206 (2002) 295

TABLES

Tab 1: Parameters settings for each level of the factorial plan.

n-C16 n-C28

Level T P WHSV H2/n-paraffin T P WHSV H2/n-paraffin

# °C bar h-1 wt/wt °C bar h-1 wt/wt

-2 290 20 1,0 0,050 285 20 1,0 0,050

-1 300 35 1,5 0,075 295 35 1,5 0,075

0 310 50 2,0 0,100 305 50 2,0 0,100

1 320 65 2,5 0,125 315 65 2,5 0,125

2 330 80 3,0 0,150 325 80 3,0 0,150

16

FIGURE CAPTIONS

Fig.1: Typical cracking products distribution for central runs when feeding n-C28 (ca.

50% conversion). PC=linear, MC=mono-branched (methyl-+ethyl-substituted),

DC=di-methyl substituted, TC=tri-methyl substituted paraffins.

Fig.2: Product distribution (%wt in the collected liquid and gas phases) as a function of

conversion for a) n-C16 and b) n-C28. “Mono-“ and “multibranched” products refer

to the feed isomerisation only, “cracking” products were calculated as C1-Cn-1

lump (linear + branched isomers).

Fig.3: Dependence of conversion of n-C16 and n-C28 on the operating parameters: a)

temperature, b) total pressure, c) H2/n-paraffin ratio and d) space velocity. The

numbers on figure 2c represent the percentage of evaporated wax.

Fig.4: Conversion as a function of a) the WHSV for n-C16 and b) the WHSVL, of the

liquid phase, for n-C16 and n-C28. The numbers on figure 4b represent the

experimental H2/n-paraffin ratio.

Fig.5: First order reaction plot for n-C16 and n-C28 conversion. (Y = - ln(1 - yf)). Contact

time (’) refers to the paraffin in liquid phase.

Fig.6: Reaction order with respect to H2. The numbers on the plot represent the degree

of evaporation of the wax.

17

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0 5 10 15 20 25 30

number of C atoms

%m

ol

DC

MC

TC

PC

Fig. 1a

Fig. 1b

18

0

5

10

15

20

25

30

35

40

45

0 10 20 30 40 50 60 70 80 90 100

n-C16 conversion (%)

pro

ducts

(w

t%)

Monobranched

Multibranched

Cracking

Fig. 2a

0

5

10

15

20

25

30

35

40

0 10 20 30 40 50 60 70 80 90 100

n-C28 conversion (%)

pro

du

cts

(%

wt)

Monobranched

Multibranched

Cracking

Fig. 2b

19

0

20

40

60

80

100

280 300 320 340

T (°C)

Co

nve

rsio

n (

mo

l%)

n-C16

n-C28

0

10

20

30

40

50

60

70

80

90

0 20 40 60 80 100

P (bar)

Co

nve

rsio

n (

mo

l%)

n-C16

n-C28

Fig. 3a Fig. 3b

18.818.317.7 17.0 16.0

75.1 81.7

86.2 89.9

93.1

40

45

50

55

0.050 0.075 0.100 0.125 0.150

H2/wax (wt/wt)

Co

nve

rsio

n (

%m

ol)

C28 C16

0

20

40

60

80

100

0 1 2 3 4

WHSV (h-1

)

Co

nve

rsio

n (

mo

l%)

n-C16

n-C28

Fig. 3c Fig. 3d

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0 0.1 0.2 0.3 0.4 0.5 0.6

1/WHSV (h)

-ln

(1

-yf)

H2/wax=0.05 wt/wt

H2/wax=0.075 wt/wt

H2/wax=0.1 wt/wt

H2/wax=0.125 wt/wt

H2/wax=0.15 wt/wt

0.15

0.1250.1

0.0750.05

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.5 1.5 2.5 3.5 4.5 5.5 6.5 7.5

1/WHSVL (h)

-ln

(1

-yf)

n-C16

nC-28

Fig. 4a Fig. 4b

20

0.00

0.40

0.80

1.20

0 5 10 15 20

' (h)

Y

C28 C16

Fig.5

12.4%

14.5%

17.7%23.0%

33.9%

77.0%

81.2%

86.2%

92.8%

100%

-5.00

-4.50

-4.00

-3.50

-3.00

2.5 3.0 3.5 4.0 4.5

ln (PH2)

ln(r

)

C28 C16

Fig. 6